Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Presentation Title15/17/2021
1 R01 HL158825-01
Acute Alcohol Consumption and Discrete Atrial Fibrillation Events
Gregory M Marcus, MD, MAS, Eric Vittinghoff, PhD, Isaac R Whitman, MD, Sean Joyce, Vivian Yang, Gregory Nah, MA, Edward P Gerstenfeld, MD, Joshua D. Moss, MD, Randall J. Lee, MD, PhD, Byron K. Lee, MD, Zian H. Tseng, MD, MAS, Vasanth Vedantham, MD, PhD, Jeffrey E Olgin, MD, Melvin M Scheinman, MD, Henry Hsia, MD, Rachel Gladstone, Shannon Fan, Emily Lee, Christina Fang, Kelsey Ogomori, BA, Robin Fatch, Judith A Hahn, PhD
The HOLIDAY (HOw ALcohol InDuces Atrial TachYarrhythmias) Monitors Study
Disclosures Research
- NIH (NIBIB, NCI, NHLBI)- PCORI- TRDRP- Medtronic- Eight Sleep - Baylis
Consulting- InCarda Therapeutics - Johnson and Johnson
Equity - InCarda Therapeutics (as co-founder)
7
Original Descriptors Focused on Acute Effects: The “Holiday Heart
Syndrome” Small case series relying on self-reported alcohol consumption
9
Could patients be mistaken?
Because alcohol is so commonly consumed, such observations are prone to an “availability heuristic,”1 where recent exposures might naturally, and yet potentially erroneously, be inferred as causal.
1. Tversky A, Kahneman D. Science 1974
10
Biological Plausibility
Atrial fibrillation patients randomly assigned to an alcohol infusion titrated to 0.08% breath alcohol concentration versus double-blinded, volume, and osmolality-matched placebo infusion.
No differences in AF inducibility were observed in this study of near-immediate alcohol-related effects.
Marcus GM et al. JACC Clin Electrophysiol 2021
11
Hypotheses
The timing and occurrence of an atrial fibrillation episode is not due to random chance alone
Acute alcohol consumption heightens the risk for a discrete atrial fibrillation episode
12
Methods
We enrolled 100 consenting paroxysmal atrial fibrillation patients at least 21 years of age who consumed at least one alcoholic drink per month.- With no plans to change management in the next 4 weeks- No evidence of substance or alcohol use disorders (using chart
review and the AUDIT-C questionnaire)
13
Methods 1 month of Wearable continuous ECG monitoring was
obtained using:- Lifewatch (Lifewatch,
Rosemount, Ill) ACT monitor in the first 27 participants
- Two successive Zio patches (iRhythm, San Francisco, CA) in the remining 73 participants
14
Methods Using these continuously worn ECG monitors:
- The timing of every AF episode lasting at least 30 seconds was determined
Participants were instructed to press a patient activator button on the ECG monitor only and every time they had a standard alcoholic drink.- Providing a time stamp for each drink- A standard drink was defined as a glass of wine, 12 ounce can/
bottle of beer, or a shot of hard liquor/ spirits
15
Methods For passive continuous alcohol monitoring, participants were fit with a transdermal
alcohol sensor placed around the ankle, the Secure Continuous Remote Alcohol Monitor (SCRAM, Highlands Ranch, CO).
- These devices achieve optimal sensitivity when ≥2 drinks are consumed on one occasion
- A “UCSF Cardiology” sticker was applied to mitigate concern regarding stigma
16
Methods Participants returned for in-person visits at two and four weeks to
assure compliance with devices and place new devices (such as the second Zio patch) if needed.
During those visits, a finger-stick blood spot was collected for phosphatidylethanol testing (PEth, United States Drug Testing Laboratories, Des Plaines, Illinois).- PEth is a byproduct of ethanol metabolism indicative of alcohol
consumption within the previous 21 days. - Optimal sensitivity is achieved with binge drinking.
17
Statistical Analyses We employed a case-crossover analysis among all participants that
exhibited at least one atrial fibrillation episode during the monitoring period.
The associations of having an AF episode with alcoholic consumption was estimated using conditional logistic models adjusting for day of week and time of day.- The exact time of the first AF episode was compared to the same time on non-
AF days- The initial “look-back” period for self-reported drinking events was set to 4
hours based on the median time estimated by participants- As the transdermal sensor ascertains alcohol levels only every 30 minutes and
may exhibit delayed effects, the look-back time for those events was set to 12 hours
18
Results
Participants wore the ECG monitor a median of 27 days (IQR 15-28), with 90% wearing it more than 21 days.
Real-time recordings of alcohol consumption revealed a median of 19 drinks (IQR 10-38) on a median 12 (IQR 7-21) different days.
56 participants exhibited at least one atrial fibrillation episode. Atrial fibrillation occurred on a median of 5 (IQR 2.5-12.5)
different days.
19
No AF episode(s)
(N=44)
AF episode(s) (N=56)
P-value
Mean Age (years) 64.8 ± 11.6 63.8 ± 16.9 0.72
Male sex 32 (72.7%) 46 (83.6%) 0.19
Race/ethnicity 0.84
White 37 (84.1%) 48 (85.7%)
Asian 4 (9.1%) 5 (8.9%)
Black 2 (4.5%) 1 (1.8%)
LatinX 0 (0%) 1 (1.8%)
Other 1 (2.3%) 2 (3.6%)
Hypertension 25 (56.8%) 23 (41.8%) 0.14
Diabetes 5 (11.4%) 6 (10.9%) 0.94
Coronary artery disease 10 (23.3%) 6 (10.9%) 0.10
Congestive heart failure 3 (6.8%) 3 (5.6%) 0.80
Smoking 0.67
Never 20 (45.5%) 30 (54.5%)
Former 23 (52.3%) 24 (43.6%)
Current 1 (2.3%) 1 (1.8%)
Antiarrhythmic Medications
Amiodarone 1 (2.3%) 0 (0.0%) 0.26
Dronedarone 1 (2.3%) 0 (0.0%) 0.26
Propafenone 2 (4.9%) 5 (9.4%) 0.40
Flecainide 12 (28.6%) 5 (9.3%) 0.014
Sotalol 0 (0.0%) 1 (2.0%) 0.37
20
Adjusting for study week and taking repeated measures within individuals into account, every additional participant activation-based drinking event was associated with a 23% greater odds of a positive PEth (OR 1.23, 95% CI 1.09-1.39, p<0.001).
The Spearman correlation between real-time recordings of alcohol consumption and daily areas under the curve for SCRAM-detected events was 0.52 (p<0.001).
21
Results: real time self-reported events
Alcohol consumption as a predictor of subsequent discrete atrial fibrillation episodes
Odds ratios are adjusted for day of the week
23
Results: transdermal alcohol sensor Every 0.1% increase in the inferred peak blood alcohol
concentration in the last 12 hours was associated with a 38% greater odds of an atrial fibrillation episode (OR 1.38, 95% CI 1.04-1.83, p=0.024).
The total area under the curve of alcohol exposure in the past 12 hours was also associated with a heightened risk for an atrial fibrillation episode:- Odds ratio of 1.14 (95% CI 1.06-1.22) per 4.7% increase in
alcohol exposure, p<0.001). No threshold effects were observed.
24
Strengths and Limitations The case-crossover mitigates against confounding as the same
individuals served as cases and controls. We cannot exclude the possibility that some concomitant behavior or
exposure that occurred with alcohol consumption was a causal factor.- Only 1 participant in the analysis smoked- Caffeine unlikely to be consumed at the same time
Poor sleep may occur due to alcohol consumption and lead to atrial fibrillation.- Poor sleep in this circumstance would then serve as a mediator (along the
causal pathway) rather than a confounder
25
Conclusions
Alcohol consumption appears to heighten the risk that a discrete atrial fibrillation event will occur.
This relationship exhibits a delayed effect of several hours. No clear or consistent threshold of alcohol required was
observed.- With evidence that even one drink may heighten the risk
26
Conclusions These data suggest that the probability a particular atrial
fibrillation event will occur is not simply due to random chance. A common behavior, alcohol consumption, is a modifiable
exposure that may empower patients to influence the risk of an atrial fibrillation event.
These findings suggest understanding other acute triggers of atrial fibrillation may be a fruitful field of investigation.